1. QoS interconnect best without effort Bob Briscoe Chief Researcher BT Sep 2009 This work is partly funded by Trilogy, a research project supported by the European Community www.trilogy-project.org www.trilogy-project.org

2. 2 © British Telecommunications plc BT’s Interoperability: Bespoke approach to customers Connect to one, connect to many BT provides transaction and clearing, security, reporting and monitoring to optimise all services on each layer BT provides a building-block approach to match all CP and SP needs from inter-provider connectivity to value-added services BT provides specific services based on customers’ needs, allowing them to pick and choose from the different layers what fits their market ambitions Complete range of IP-based services relying on BT 21CN and BT Global IPX in an open and flexible environment where CPs and SPs can seamlessly select BT IP Interoperability services adapted to their needs and benefit from the IP convergence world 21CN Framework (network, platform, application, etc.)

3. 3 © British Telecommunications plc Framework (network, platform, application, etc.) 21CN BT's Interoperability – IP Exchange (IPX) PSTN network Mobile network CATV IP Voice network DSL IP Voice network IP Exchange Call Manager Billing & Charging Number Translation Management Information IP SS7/IP IP Exchange is at the heart of BT‘s Interoperability portfolio. IP Exchange enables communication between different infrastructures handling voice and data traffic to ensure that users can communicate to each other irrespective of – the service – the location – the device or network IP Exchange ensures that all parties involved in a service usage can be billed and charged appropriately

4. QoS interconnect best without effort coming soon…?

5. © British Telecommunications plc both value and cost industry contractual metrics are largely value-based e.g. volume ratio, session instances even a CEO should understand both value and cost competitive market drives revenues down towards provider’s marginal cost those who understand marginal costs will succeed time consumer value provider cost cost to consumer = provider revenue consumer surplus provider profit Vol i Vol o

6. © British Telecommunications plc simplicity ahead! cannot be QoS on exit check mirrors – it was QoS

7. © British Telecommunications plc the big idea pool all traffic together real-time, p2p file-sharing, Web, streaming, … keep heavy traffic from harming QoS sensitive apps through economic incentives backed by enforcement once cost-based mechanisms correct for all traffic serves as cost-based floor for value-based services

8. © British Telecommunications plc congestion is not evil congestion signals are healthy no congestion across whole path is evil –for data transfer to complete ASAP, must fill bottlenecks the trick signal congestion just before impairment –explicit congestion notification (ECN) 2001 update to IP: as a queue builds mark more packets –then tiny queuing delay and tiny tiny loss for all traffic time bit- rate time bit- rate 

9. © British Telecommunications plc marginal cost of network usage? volume is NOT a good measure heavy user yields at early congestion onset very high volume but very low cost to others e.g. LEDBAT (BitTorrent’s or Microsoft’s low extra delay background transport) or weighted TCP by counting volume, operators kill nice behaviour correct measure: congestion-volume your contribution to congestion = bytes marked bit-rate time congestion time 10GB 0.01% marking 1MB 1% marking 1MB 300MB 100MB 3MB

10. © British Telecommunications plc 'Cost': Congestion-Volume: Total TCP Loss [Byte] Initial results measured on Naples Uni net Each point is a user correlation coefficient: 0.43 Volume: Total TCP Traffic Volume [Byte] 100% WARNING: Requires validation with more sample data  volume  congestion-volume user’s contribution to congestion

11. © British Telecommunications plc congestion-volume metric dual demand & supply role a resource accountability metric 1.of customers to ISPs (too much traffic) 2.and ISPs to customers (too little capacity) 1.cost to other users of my traffic 2.the marginal cost of upgrading equipment so it wouldn’t have been congested competitive market matches 1 & 2 note: diagram is conceptual congestion volume would be accumulated over time capital cost of equipment would be depreciated over time

12. © British Telecommunications plc congestion exposure by Internet design, endpoints detect & handle congestion v hard for networks to see losses (the marginal cost) explicit congestion notification (ECN) not enough visible, but too late – as packets leave the network proposed IETF working group: “congestion exposure” re-ECN: sender marks IP headers to expose congestion to measure traffic cost as easily as we measure volume just count volume of marked packets in aggregate >40 significant offers of help just in last fortnight Microsoft, Nokia, Cisco, Huawei, Alcatel-Lucent, NEC, Ericsson, NSN, Sandvine, Comcast, DT, Verizon, …

13. © British Telecommunications plc incentive for apps to avoid congestion backed by enforcement example consumer use of exposed congestion still with flat fee bulk congestion policer Internet 0.3% congestion 0% 0.1% 2 Mb/s 0.3Mb/s 6 Mb/s Acceptable Use Policy 'congestion-volume' allowance: 1GB/month @ €15 / month Allows ~70GB per day of data in typical conditions

14. © British Telecommunications plc NDND NANA NBNB NCNC solution legend: reveals bulk marginal cost ‘routing money’ re-ECN downstream congestion marking [%] bit rate area = instantaneous downstream congestion volume just two counters at border meter monthly bulk congestion-volume = true marginal cost without measuring flows 0|0|2|7|6|0|5 € € €

15. © British Telecommunications plc I’m a conservative, get me out of here! if we don’t listen to the economics, we’re all dead shift from value-based to cost-based is unstoppable –competition bit transport needs to be viable on its own (another talk) as cost pressures grow existing capacity sharing methods feed an arms race –TCP doesn’t share capacity fairly by any means recent unanimous consensus in IETF Transport Area –ISPs have quietly been fighting TCP with piecemeal tools WFQ, volume capping, deep packet inspection with congestion in IP header, wouldn’t need to look deeper

16. © British Telecommunications plc guaranteed bit-rate? or much faster 99.9% of the time? harnessing flexibility the idea that humans want to buy a known fixed bit-rate –comes from the needs of media delivery technology –hardly ever a human need or desire services want freedom & flexibility –access to a large shared pool, not a pipe when freedoms collide, congestion results –many services can adapt to congestion –shift around resource pool in time/space constant quality video encoding Constant Bit Rate 100%Constant Quality 125% sequences encoded at same average of 500kb/s Equitable Quality 216% [Crabtree09] % figures = no. of videos that fit into the same capacity time bit rate

17. © British Telecommunications plc 1995 2009 telco /NGN Internet cellular satellite cable bringing cost information to the control point Internet no control without information re-ECN packets reveal real-time cost flat fee policer was just one example... huge space for business & technical innovation at the control point cost based, value-cost based bulk, per flow, per session call admission control policing, charging tiers, continuous wholesale, retail truly converged architecture can apply different industry cultures through policies at the control point not embedded in each technology

18. © British Telecommunications plc best without effort did you notice the interconnected QoS mechanism? –endpoints ensure tiny queuing delay & loss for all traffic –if your app wants more bit-rate, it just goes faster –visible in bulk metric at every border (for SLAs, AUPs) simple – and all the right support for operations the invisible hand of the market –favours ISPs that get their customers to manage their traffic in everyone else‘s best interests incentives to cooperate across Internet value chain –content industry, CDNs, app & OS authors, network wholesalers & retailers, Internet companies, end-customers, business, residential

19. © British Telecommunications plc more info... BT IP Exchange jens.specht@bt.com White paper – the whole story in 7pp Internet: Fairer is Faster, Bob Briscoe (BT), BT White Paper TR-CXR9-2009-001 (May 2009) - an abridged version of this article appeared in IEEE Spectrum, Dec 2008Internet: Fairer is Faster Inevitability of policing The Broadband Incentives Problem, Broadband Working Group, MIT, BT, Cisco, Comcast, Deutsche Telekom / T-Mobile, France Telecom, Intel, Motorola, Nokia, Nortel (May ’05 & follow-up Jul ’06) cfp.mit.edu Stats on p2p usage across 7 Japanese ISPs with high FTTH penetration Kenjiro Cho et al, "The Impact and Implications of the Growth in Residential User-to-User Traffic", In Proc ACM SIGCOMM (Oct ’06) How wrong Internet capacity sharing is and why it's causing an arms race Bob Briscoe et al, "Problem Statement: Transport Protocols Don't Have To Do Fairness", IETF Internet Draft (Jul 2008)Problem Statement: Transport Protocols Don't Have To Do Fairness Understanding why QoS interconnect is better understood as a congestion issue Bob Briscoe and Steve Rudkin "Commercial Models for IP Quality of Service Interconnect" BT Technology Journal 23 (2) pp. 171--195 (April, 2005)Commercial Models for IP Quality of Service Interconnect Re-architecting the Internet: The Trilogy projectTrilogy Congestion Exposure (re-ECN) at the IETF trac.tools.ietf.org/area/tsv/trac/wiki/re-ECN

20. QoS interconnection best without effort Q&A...

21. © British Telecommunications plc problems using congestion in contracts 1.loss: used to signal congestion since the Internet's inception computers detect congestion by detecting gaps in the sequence of packets computers can hide these gaps from the network with encryption 2.explicit congestion notification (ECN): standardised into TCP/IP in 2001 approaching congestion, a link marks an increasing fraction of packets implemented in Windows Vista (but off by default) and Linux, and IP routers (off by default) 3.re-inserted ECN (re-ECN): standards proposal since 2005 packet delivery conditional on sender declaring expected congestion uses ECN equipment in the network unchanged 1. loss2. ECN3. re-ECN can't justify selling an impairment  absence of packets is not a contractible metric  congestion is outside a customer's control  customers don't like variable charges  congestion is not an intuitive contractual metric 

22. © British Telecommunications plc packet headers data 1 probability drop mark ave queue length ACKnowledgement packets network transport data probabilistic packet marking algorithm on all egress interfaces marked packet marked ACK explicit congestion notification (ECN) ECN bits 6 & 7 of IP DS byte 00:Not ECN Capable Transport (ECT) 01 or 10:ECN Capable Transport - no Congestion Experienced (sender initialises) 11:ECN Capable Transport - and Congestion Experienced (CE) DSCP 0 5 6 7 IETF proposed std: RFC3168 Sep 2001 most recent change to IPv4&6 IETF proposed std: RFC3168 Sep 2001 most recent change to IPv4&6

23. © British Telecommunications plc Diff serv ECNECN RERE IPv4 header Feedback path Data packet flow Sender Receiver +1+1 Routers Networks 1. Congested queue debit marks some packets 2. Receiver feeds back debit marks 3. Sender re-inserts feedback (re-feedback) into the forward data flow as credit marks 4. Outcome: End-points still do congestion control But sender has to reveal congestion it will cause Then networks can limit excessive congestion 5. Cheaters will be persistently in debt So network can discard their packets (In this diagram no-one is cheating) 1 2 3 5 4 congestion exposure in one bit standard ECN (explicit congestion notification + re-inserted feedback (re-feedback) = re-ECN no changes required to IP data forwarding

24. © British Telecommunications plc main steps to deploy re-feedback / re-ECN network turn on explicit congestion notification in routers (already available) deploy simple active policing functions at customer interfaces around participating networks passive metering functions at inter-domain borders terminal devices (minor) addition to TCP/IP stack of sending device or sender proxy in network customer contracts include congestion cap oh, and first we have to update the IP standard started process in Autumn 2005 using last available bit in the IPv4 packet header proposal for new working group, Nov 2009 IETF